SELECTION IN BAR-EYED RACE OF DROSOPHILA. 387 



sterile and the other six produced offspring. In each case the 

 specimens were mated with bar-eyed flies and with one exception 

 the full-eyed males produced nothing but bar-eyed males and 

 heterozygous females and the heterozygous females produced 

 bar-eyed and full-eyed males and bar-eyed and heterozygous 

 females. The one exception occurred in case of the male which 

 appeared on March 28 and was mated to a bar-eyed female 

 on April 3. The bottle was set aside and overlooked until 

 April 27 when it was found to contain all four classes of flies as 

 shown in the table. In that case the full-eyed males and bar- 

 eyed females are undoubtedly due to the appearance of flies of 

 the second generation before the count was made. Since the 

 first generation consisted of bar-eyed males and heterozygous 

 females the second generation would be expected to contain a 

 mixture of all four classes. 



The facets in the eyes of all the offspring obtained from the 

 male appearing on November 15 were counted and the results 

 are given under VBax in Table I. and Fig. 2. A large number 

 of pairs from these offspring was mated, but all proved to be 

 sterile. A mass mating made toward the end of the hatching 

 period, however, produced 14 bar-eyed males with an average 

 facet number of 114, 7 full-eyed males, 12 bar-eyed females with 

 an average facet number of 91 and 5 heterozygous lemales. 

 These were bred in a stock bottle for several generations to 

 see if one or the other class would become dominant, but no 

 obvious change in the ratios appeared. 



DISCUSSION AND CONCLUSIONS. 



The variability of the stock at the beginning of the selections 

 made it appear probable that the facet number is affected by 

 environmental conditions. For that reason the records during 

 the experiments were kept more carefully than would otherwise 

 have been the case. Counts were made and recorded separately 

 at least every twelve hours. Each line or branch of a line was 

 recorded separately, and in cases where parents were trans- 

 ferred and offspring obtained from both bottles a separate record 

 was kept of each. Observations were also made on a possible 

 correlation between bodv size and facet number. 



